What it does

mention-intel takes a handle, brand, ticker, or phrase and runs the workflow agents already stitch together by hand: search for mentions, then enrich the authors of the mentions that matter, then read the overall tone. 

{
  "query": "agentutility.ai",
  "max_results": 10
}

returns the mentions with each author's follower count and verification status, an authenticity score for the top authors, and a sentiment breakdown across the mention text: 

{
  "mentions": [
    {
      "text": "agentutility.ai is shipping x402 endpoints fast",
      "url": "https://x.com/agentutility_ai/status/1234567890",
      "author": { "handle": "agentutility_ai", "followers": 4200, "verified": false }
    }
  ],
  "sentiment": { "overall": "positive", "score": 0.4 },
  "top_authors": [{ "handle": "agentutility_ai", "followers": 4200, "authenticity_score": 78 }]
}

Why

Mentions search (mentions, tweet-search) sees real call volume, and the follow-up step — enriching the authors of the mentions worth paying attention to, then checking whether the account is real — is a second and third call agents were already running against twitter-user-lookup and x-account-authenticity. Collapsing that into one call means one settlement instead of three, and the enrichment only runs against the top five distinct authors in the result, not every mention.

Degradation

mentions is the required leg — if the search fails the call fails. Per-author enrichment and the batch sentiment pass are independent optional legs: an X API rate limit or an LLM provider hiccup degrades that part of the response (author fields fall back to the values already present in the search result; sentiment reports null with a note) instead of failing the whole call.

Price: $0.06.